Back to Documentation
DocsCLI Referenceyamlr drift
Last Updated: April 2026

yamlr drift

Live-Manifest Reconciliation Engine (PRO)

The yamlr drift command audits your live Kubernetes cluster against your local YAML manifests to detect undocumented changes and out-of-band configurations.

Usage

bash
# Basic Drift Detection (All Namespaces) yamlr drift . # Filter by Specific Namespace yamlr drift . --namespace production # Target a specific Kubeconfig context yamlr drift . --context eks-cluster-1

How it works

Yamlr analyzes the "Live Truth" from your cluster and compares it with the "Proposed Truth" in your local directory using its Semantic DNA engine. It ignores non-vital mutations (like status timestamps or system annotations) and focuses on structural drift (image tags, replicas, resource limits).

Enterprise Features

1. Context Isolation (--context)

Explicitly target a cluster defined in your ~/.kube/config. This is essential for preventing "Drift Leakage" when managing multiple environments.

2. Namespace Scoping (--namespace)

Limit the audit to a specific namespace to reduce noise and analysis latency in multi-tenant clusters.

3. Reporting Tiers

  • Missing: Local manifests that are not found in the cluster.
  • Shadow: Live resources that have no local manifest counterpart.
  • Mutated: Structural differences between local and live state.

[!IMPORTANT] yamlr drift is a Reporting-Only operation in the current version. It does not automatically "heal" the cluster back to the local state. To apply fixes, use the scan and heal workflow.

[!NOTE] This feature requires a [PRO] license. Run yamlr auth login to activate your enterprise tier.